Debian Bug report logs - #639539
load average reported incorrectly low

Reply or subscribe to this bug.

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Bart Bobrowski <mrbart@gmail.com>

To: submit@bugs.debian.org

Subject: load average reported incorrectly low

Date: Sun, 28 Aug 2011 00:10:59 +0200

Subject: base: load average reported incorrectly low
Package: base
Severity: important

*** Please type your report below this line ***
EVER SINCE UPGRADING TO SQUEEZE, LOAD AVERAGE IN TOP IS INCORRECTLY
REPORTED AS TOO LOW THAN IT REALLY IS. THIS PROBLEM IS PRESENT IN THE
DEFAULT SQUEEZE KERNEL, AS WELL AS UPGRADED 6.38 KERNEL. THIS PROBLEM
DOES NOT EXIST IN 2.6.26-2-amd64 LENNY. THIS PROBLEM HAS BEEN
REPLICATED ON MORE THAN 1 SERVER. DOWNGRADING TO LENNY KERNEL FIXES
THE PROBLEM. EXAMPLE: SERVER ON LENNY RUNNING 0.80 0.70 0.60 LOAD
AVERAGE LATER REGISTERS 0.00 0.02 0.05 AFTER UPGRADE DURING THE SAME
SERVER LOAD THROUGHOUT THE DAY. SERVER IS RUNNING MYSQL, PROBLEM ALSO
EXISTS RUNNING PHP WEB SERVER. PROBLEM HOWEVER DOES NOT SEEM PRESENT
DOING OTHER TASKS, SUCH AS COMPILING OR GZIPPING.

-- System Information:
Debian Release: 6.0.2
  APT prefers stable-updates
  APT policy: (500, 'stable-updates'), (500, 'stable')
Architecture: amd64 (x86_64)

Kernel: Linux 2.6.38-bpo.2-amd64 (SMP w/4 CPU cores)
Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1)
Shell: /bin/sh linked to /bin/dash

Message #10 received at 639539@bugs.debian.org (full text, mbox, reply):

From: Holger Levsen <holger@layer-acht.org>

To: 639539@bugs.debian.org

Subject: Re: Bug#639539: load average reported incorrectly low

Date: Sun, 28 Aug 2011 08:30:57 +0200

my eyes!

Message #15 received at 639539@bugs.debian.org (full text, mbox, reply):

From: Bart Bobrowski <mrbart@gmail.com>

To: 639539@bugs.debian.org

Subject: Re: Bug#639539: Acknowledgement (load average reported incorrectly low)

Date: Thu, 1 Sep 2011 00:39:43 +0200

a video of the problem: http://www.youtube.com/watch?v=KjSZ-D0XBBQ

-bart



On Sun, Aug 28, 2011 at 12:18 AM, Debian Bug Tracking System
<owner@bugs.debian.org> wrote:
> Thank you for filing a new Bug report with Debian.
>
> This is an automatically generated reply to let you know your message
> has been received.
>
> Your message is being forwarded to the package maintainers and other
> interested parties for their attention; they will reply in due course.
>
> Your message has been sent to the package maintainer(s):
>  Base Maintainers <virtual-pkg-base-maintainers@lists.alioth.debian.org>
>
> If you wish to submit further information on this problem, please
> send it to 639539@bugs.debian.org.
>
> Please do not send mail to owner@bugs.debian.org unless you wish
> to report a problem with the Bug-tracking system.
>
> --
> 639539: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=639539
> Debian Bug Tracking System
> Contact owner@bugs.debian.org with problems
>

Message #20 received at 639539@bugs.debian.org (full text, mbox, reply):

From: Amaya <amaya@debian.org>

To: Bart Bobrowski <mrbart@gmail.com>, 639539@bugs.debian.org

Cc: debian-backports@lists.debian.org

Subject: Re: [Virtual-pkg-base-maintainers] Bug#639539: Acknowledgement (load average reported incorrectly low)

Date: Mon, 5 Sep 2011 12:59:30 +0200

The original bug report is at http://bugs.debian.org/639539 (please
excuse the capitalization).


Bart Bobrowski wrote:
> a video of the problem: http://www.youtube.com/watch?v=KjSZ-D0XBBQ

You seem to be using a kernel form backports, so we need to reassign
this bug report there.



-- 
 .''`.  Sell Crazy Someplace Else. We're all stocked up here
: :' :
`. `'
  `-                       Proudly running Debian GNU/Linux

Message #25 received at 639539@bugs.debian.org (full text, mbox, reply):

From: Bart Bobrowski <mrbart@gmail.com>

To: 639539@bugs.debian.org

Subject: Re: Bug#639539: Acknowledgement (load average reported incorrectly low)

Date: Mon, 5 Sep 2011 14:24:48 +0200

Same problem with the default squeeze installation... so 2.6.32, I was
hoping 6.38 would clear things up but no...

-bart



On Sun, Aug 28, 2011 at 12:18 AM, Debian Bug Tracking System
<owner@bugs.debian.org> wrote:
> Thank you for filing a new Bug report with Debian.
>
> This is an automatically generated reply to let you know your message
> has been received.
>
> Your message is being forwarded to the package maintainers and other
> interested parties for their attention; they will reply in due course.
>
> Your message has been sent to the package maintainer(s):
>  Base Maintainers <virtual-pkg-base-maintainers@lists.alioth.debian.org>
>
> If you wish to submit further information on this problem, please
> send it to 639539@bugs.debian.org.
>
> Please do not send mail to owner@bugs.debian.org unless you wish
> to report a problem with the Bug-tracking system.
>
> --
> 639539: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=639539
> Debian Bug Tracking System
> Contact owner@bugs.debian.org with problems
>

Message #30 received at 639539@bugs.debian.org (full text, mbox, reply):

From: Rob Donovan <hikerman2005-hjkdfgfdgngffdcvbb@yahoo.com>

To: 639539@bugs.debian.org

Subject: Load average bug

Date: Wed, 07 Sep 2011 11:25:46 -0700

I can confirm that I'm seeing this bug too after an upgrade to the 
standard Debian Squeeze distribution from Lenny (new kernel is 
2.6.32-5-amd64).

It shows up in top and, of course, xload.

In my case the load average comes in at about 0.3 for a nice-19 job 
eating 784% of my 8-cpu processor (max being 800%).
Under Lenny xload displayed similar jobs as (almost) 8 bars.  Now it's 
much less than 1 bar.

xload is therefore effectively broken.

Rob Donovan

Message #35 received at 639539@bugs.debian.org (full text, mbox, reply):

From: Rob Donovan <hikerman2005-hjkdfgfdgngffdcvbb@yahoo.com>

To: 639539@bugs.debian.org

Subject: ps

Date: Wed, 07 Sep 2011 11:49:58 -0700

This may be related :

I run jobs that eat nearly 800% of my 800% available CPU that take about 
10 days to complete, and I run them at nice 19.  Under Lenny such jobs 
always showed up at the top of top reporting 700% + CPU consumption, as 
you'd expect.  I upgraded to Squeeze on 19 August 2011, and since then 
such jobs have still shown up at the top of top... with one exception.  
On one occasion the job was still running (I could see it logging to 
stdout at the usual speed in the window in which it was running) but it 
wasn't at the top of top.  In fact, only by restricting top to show 
processes owned by me (u <userid>) could I see it at all, and it was 
then displayed at the bottom of my processes.  Memory was displayed OK, 
but CPU consumption... was displayed as zero I think.  I can't remember 
what the CPU line at the top of the screen was displaying.  If it 
happens again I'll make notes.

Message #40 received at 639539@bugs.debian.org (full text, mbox, reply):

From: Rob Donovan <hikerman2005-hjkdfgfdgngffdcvbb@yahoo.com>

To: 639539@bugs.debian.org

Subject: top problem

Date: Mon, 12 Sep 2011 22:57:30 -0700

has happened again :

top - 22:49:30 up 7 days,  7:38,  3 users,  load average: 1.19, 0.88, 0.80
Tasks: 212 total,   3 running, 209 sleeping,   0 stopped,   0 zombie
Cpu(s):  0.0%us,  1.5%sy, 92.7%ni,  5.9%id,  0.0%wa,  0.0%hi,  0.0%si,  
0.0%st
Mem:  16461424k total, 11328664k used,  5132760k free,   484068k buffers
Swap: 35150104k total,        0k used, 35150104k free,  8270712k cached

 PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
5178 user       20   0 19180 1436 1004 R    0  0.0   0:00.19 top
3431 user       20   0 79296 1904  944 S    0  0.0   1:53.73 sshd
3432 user       20   0 21720 4592 1620 S    0  0.0   0:00.17 bash
3474 user       20   0  186m  26m  12m S    0  0.2   5:26.99 emacs
3478 user       20   0 26104  928  624 S    0  0.0   0:00.00 dbus-launch
3479 user       20   0 23260  824  580 S    0  0.0   0:00.00 dbus-daemon
3481 user       20   0 46008 5548 2316 S    0  0.0   0:00.44 gconfd-2
3482 user       39  19 1844m 1.6g  932 S    0 10.1  2590446h my_prog.x
3488 user       20   0 79316 2044  960 S    0  0.0   0:02.31 sshd
3489 user       20   0 21764 4640 1620 S    0  0.0   0:00.60 bash
16039 user       20   0 78988 1724  956 S    0  0.0   0:00.12 sshd
16216 user       20   0 21720 4612 1632 S    0  0.0   0:00.36 bash

This is just processes owned by me (userid changed to user).  The 
program that is eating all the CPU is my_prog.x, which is running nice 
19.  It's been running for about 67 hours now, not 2590446 hours.  The 
92% nice CPU usage (up the top) is credible, while 0% is not.  1844m 
memory usage is correct.  The program is logging to stdout at the usual 
rate.  Load average should be about 7.8.  CPU usage for my_prog should 
be about 760% +.  

Send a report that this bug log contains spam.